¹ This growth rate was calculated on the basis of the aggregate population data collected by the Bakufu and cited in Naotarō, Sekiyama, Kinsei Nihon no jinkō kōzō (The Population Structure in Tokugawa Japan) (Tokyo: Yoshikawa Kōbunkan, 1958), p. 139Google Scholar.

² Eijirō's, Honjō “The Population of Japan in the ‘Tokugawa’ Era,” Tokugawa Bakufu no beika chōsetsu (Price Adjustments by the Tokugawa Bakufu) (Tokyo: Kōbundō Shobō, 1924)Google Scholar probably provides the best summary in English of this view. See also Taeuber, Irene B., The Population of Japan (Princeton: Princeton University Press, 1958)Google Scholar , Chapter 2 for a summary of the views of the major Japanese scholars on this subject. The best known exponents of this view in Japanese are Takahashi Bonsen and Honjō Eijirō. Takahashi has concentrated on the practices and effects of abortion and infanticide in Tokugawa Japan. See his Datai mabiki no kenkyū (A Study of Abortion and Infanticide) (Tokyo: Chūō Shakai Jigyō Kenkyūsho, 1936)Google Scholar and his Nihon jinkōshi no kenkyū (A Study of Japan's Population History), Vol. 1 (Tokyo: Sanyūsha, 194Google Scholar :), Vol. 2 (Tokyo: Nihon Gakujutsu Shinkōkai, 1955), and Vol. 3 (Tokyo: Nihon Gakujutsu Shinkōkai, 1962).

³ The major Japanese scholar currently studying the Tokugawa population is Hayami Akira of Keiō University who has published his research results in a number of articles in Keiō's journal, Mita Gakkai Zasshi (Mita Journal of Economics).

⁴ For evidence in English, see Smith, Thomas C., “The Land Tax in the Tokugawa Period,” Journal of Asian Studies, XVIII, No. 1 (Nov. 1958), pp. 3–19CrossRefGoogle Scholar . Writings revealing such evidence in Japanese are Hiroshi, Shimpō, Hōkenteki shōnōmin no bunkai katei (The Process of the Dissolution of Feudalistic Petty Farmers) (Tokyo: Shinseisha, 1967)Google Scholar; Jun'nosuke, Sasaki, Bakuhan kenryoku no kiso kōzō: “shōnō” jiritsu to gunyaku (The Basic Structure of the Shogunate-Domain Power Structure: The Independence of the “Petty Farmers” and the Military Obligations [of the Samurai]) (Tokyo: Chuo Koronsha, 1966)Google Scholar ; and Jirō, Naitō, Honbyakushō taisei no kenkyū (A Study of the Honbyakushō System) (Tokyo: Ochanomizu Shobō, 1968)Google Scholar . For additional references, see Hanley, Susan B. and Yamamura, Kozo, “A Quiet Transformation in Tokugawa Economic History,” Journal of Asian Studies, XXX, No. 2 (Feb. 1971), pp. 373–384CrossRefGoogle Scholar . Yamamura, Kozo also provides both direct and indirect evidence in “The Increasing Poverty of the Samurai in Tokugawa Japan, 1600–1868,” The Journal of Economic History, XXXI, No. 2 (June 1971), pp. 378–406CrossRefGoogle Scholar.

⁵ The shūmon-aratame, investigation of religion, was begun in the seventeenth century as a means of controlling Christianity and preventing its dissemination throughout the country. This investigation is known to have been conducted as early as 1624, but it was not carried out on a nationwide basis until after the Shimabara Rebellion of 1637–38. Extant records date from the 1670's. The form of the shūmon-aratame-chō or registers became almost identical to that of population surveys carried out on regional basis earlier in the century. This enabled the shūmon-aratame-chō to become the basis for national population surveys, the first of which was undertaken in 1721.

How the religious investigations were carried out varied by area, and the method obviously affected the reliability. In Okayama, a copy was sent each year to the domain government and a working copy was kept in the village. The working copy was kept up to date as a record of vital statistics by adding information regarding demographic events occurring during the year. Thus deaths were noted by pasting a piece of paper (harigami) with the pertinent information to the entry of the individual who died, and persons who were born or married into a family were added to the end of the family entry. Every family entry of even the working copies has the seal of the householder verifying the contents, and thus the information can be considered to have been checked by the head each year.

Since it is the working copies left in the village which have been preserved in Fujito and in most villages, each register contains both a basic census of the village, on the first day of the year for Fujito, and a vital statistics record of demographic events which occurred during the calendar year following the census. For a description of the contents of these registers, see the following section in the text. For further discussions of the shūmon-aratame-chō and the reliability of the information contained in the registers, see Sekiyama, , Kinsei Nihon no jinkō kōzō, Chapter 1, the most complete of the sources, and also Hayami Akira, “The Demographic Analysis of a Village in Tokugawa Japan: Kando-shinden of Owari Province, 1778–1871,” Keiō Economic Studies, Vol. 5 (1968), pp. 50–88Google Scholar; Akira, Hayami, “Tokugawa-kōki Owari ichi nōson no jinkō tōkei—Kaisai-gun Kandoshinden no shūmon-aratame-chō bunseki” (Population Statistics of One Farm Village in Owari in the Latter Half of the Tokugawa Period: An Analysis of the Religious Investigation Registers of Kando-shinden of Kaisai District), Mita Gakkai Zasshi, Vol. 59, No. 1 (1966), pp. 58–77Google Scholar; Honjō, , “The Population of Japan in the ‘Tokugawa’ Era”; and Nomura Kanetarō, On Cultural Conditions Affecting Population Trends in Japan (Tokyo: The Science Council of Japan, Division of Economics and Commerce, 1953)Google Scholar.

⁶ The domain of Okayama, part of the present prefecture of Okayama, was located on Honshū facing the Inland Sea.

⁷ The records of the Hikasa family and of the village of Fujito are now located in the Okayama University Library, Okayama, Japan. The case study presented in this article is based primarily on these documents.

⁸ Shūmon-aratame-chō are extant for Fujito for the following years: 1775, 1778 (partial), 1794, 1797–1806, 1808–1810, 1825–1835, 1837, 1841, 1844–1848, 1850, 1852, 1853 (partial), 1856, 1857, 1859, 1861, and 1863.

⁹ Only 1.9 percent (12 out of 625) of the children who appeared in the Fujito registers and whose parents were married and living in the village in the year the children were born were not registered in the first year of life. In many domains it was common to register children only after they reached the second or third year of life, and in Wakayama, for example, children were not registered until age eight. Sekiyama, Kinsei Nihon no jinkō kōzō, p. 58.

¹⁰ Tests performed to check the accuracy of the Fujito data include an examination of the sex ratios, the age structure, and the distribution of births and deaths by day of month. The data were also checked for the sudden inclusion of persons previously unlisted and the omission of inhabitants from the records without stated cause. Obvious errors were corrected before the data were analyzed. The major weakness appeared to be that a number of marriages went unreported after 1840. After this date, a number of children were reported born to men who had no wives; at the same time the number of unmarried women rose. Since it is presumed that there were more marriages during this period than were reported, no attempt was made to calculate the proportions married.

¹¹ A linear regression equation in the form of Y = a + bX was calculated to show the trend in the population growdi from 1775 to 1863. The constant a was 530.15 and the regression coefficient b was 1.66. This means that the trend line started from a base population of 530.15 persons and increased at a rate of 1.66 persons per year. The coefficient of determination R ² was .7339, which means that the fit of the trend line to the actual data was good, and statistically significant at the .01 percent level. The standard error of the constant was 3.63 and that of the regression coefficient .162. This trend line indicates that the population of Fujito was growing over time at a very slow rate.

¹² The term household is here used to denote all persons included in one listing in the shūmonaratame-chō. The term family is used to denote all persons related to the head of the household. For all practical purposes, the family in Fujito is equal to the number of persons in the household minus any household servants or employees.

¹³ The relatively constant mean in household size over the period was in large part due to the small fluctuation in the number of family members around the mean in each household. In short, a family in which the modal and/or mean number was four tended to return to this number whenever the number of members either exceeded or fell below four. This was true for families of whatever size. This tendency was statistically tested by calculating the coefficient of variation for each household over time.

¹⁴ The years examined were 1775, 1794, 1802, 1810, 1825, 1833, 1837, 1844, 1856, and 1863.

¹⁵ The crude birth rates for Fujito follow the usual definition of the “ratio of total registered live births to the total population.” Barclay, George W., Techniques of Population Analysis (N. Y.: John Wiley & Sons, Inc., 1958), p. 168Google Scholar . The key, of course, is “registered,” and registration tends to vary by country and period. In the case of Fujito, no child was registered who lived less than three weeks. Obviously, children who died at birth or in early infancy were not reported, but these births can be considered unimportant in a study of the social and economic significance of changes in the population. Although both the birth and death rates of Fujito are understated by the omission of children who died in early infancy, nevertheless the birth and death rates for Fujito can be considered low. For example, see the pre-modern crude birth and death rates for England and Wales cited in Glass, D. V., “Population and Population Movements in England and Wales, 1700–1850,” in Glass, D. V. and Eversley, D. E. C., ed., Population in History (London: Edward Arnold, Ltd., 1965), pp. 221–246Google Scholar ; the rates for Scandanavia in Gustaf Utterstrom, “Two Essays on Population in Eighteenth-Century Scandinavia,” Population in History, p. 538; and rates for Italy in Carlo M. Cipolla, “Four Centuries of Italian Demographic Development,” Population in History, pp. 576–578.

¹⁶ In every case but one in which a woman was reported as having a child at age 45 or above, there was a daughter or other female of childbearing age in the family. In nearly half of the cases in which a woman over 40 was reported as giving birth, there was another woman in the household. Because of the probability that at least in some of the cases the daughter or other member had in fact given birth, probably to an illegitimate child, the birth was attributed to the reported mother only if she was under the age of 45, unless there was no other female in the family. However, the number of births attributed to women over 45 was so few that none of the calculations made would be substantially affected by whatever categorization was made.

¹⁷ To test for any change in the average age at marriage, the data were grouped by years to approximate five-year averages. The groupings were 1775–78, 1794–99, 1800–04, 1805–10, 1825–29, 1830–34, 1835–41, 1844–48, 1850–57, and 1859–63. The sample sizes ranged from 20 to 42, size being related to the number of years for which there exist data in each period. From 1794 on there was no trend in average age at marriage, and the mean fluctuated around age 24.

¹⁸ Cox, Peter R., Demography (Cambridge: Cambridge University Press, 1970), p. 152Google Scholar.

¹⁹ In any year at least one or two persons were listed who did not subsequently appear in the registers. In cases in which the register for the following year was missing, there was no way to determine whether the person was later recorded as dying or leaving the village, but in any case, this number was very small for most years. The number of such persons was significant only during the period under discussion. Since nearly all persons dropped from the records were adult males, it is likely that they left the village to work elsewhere without official permission and hence the village officials could list no official change in residence. A very few of these individuals reappeared again years later.

²⁰ The statement of the district head of Tsuu is quoted in Usaburō, Nagayama, Kurashiki shishi (The History of Kurashiki City) (Kurashiki: Kurashiki Shishi Kankō Iinkai, 1963), Vol. 7, p. 377Google Scholar.

²¹ The descriptions of the famine are from the Tempō nendo kikin jōkyo torishirabe-chō (A report on the Investigations of Famine Conditions in the Tempō Years), compiled by the Okayama District Administration in 1888 and now located in the Kurashiki Municipal Library, Kurashiki, Okayama.

²² This increase was partly due to the failure to report some marriages and partly due to an increase in widows and widowers. After the Tempō famine, an obvious attempt was made by the villagers to create “whole” families and to care for persons left without support or their immediate families. Thus widowed cousins and other peripheral relatives were taken in along with their children. Even old women were “adopted” in the 1840's.

²³ Yū, Fujikawa, Nihon shitsubyōshi (A History of Disease in Japan) (Tokyo: Heibonsha, 1969), p. 62Google Scholar.

²⁴ For example, Takahashi Bonscn focuses on the suffering of Tōhoku peasants in his Datai mabiki no kenkyū, providing both qualitative and quantitative evidence of the effects of the various famines in this region.

²⁵ The number of persons leaving the village for employment and coming into it for employment tended to follow a similar pattern, except that the number of persons leaving the village for employment was slightly larger than the number entering. For example, those leaving numbered in the thirties in 1775 and in the 1790's through the early nineteenth century, when the number began to drop. By 1810 there were only 17. In 1825, those leaving numbered only 11, and the figure continued to drop, reaching low points of four in 1837 and in 1847. Only in the 1860's did the number rise to as high as ten again.

²⁶ For example, in Kando-shinden of Owari Province, the average number of children born to women whose marriages lasted to age 45 was 7.5 if the women married between 16–20, 6.2 if they married between 21–25, and 4.8 if 26 or above. Hayami, “The Demographic Analysis of a Village in Tokugawa Japan,” pp. 57–59, 73. On the other hand, in Kōmi village of Mino Province, the average number of children living with each couple in each year examined was slightly over 2 from 1674–1872. See Kenkyūkai, Nomura, Han, Komi Mura Kyōdō Kenkyū (The Nomura Research Group for the Study of Kōmi Village), “Ogaki hanryō Mino no kuni Motosu no kōri Kōmi mura no kokō tōkei,” (Household and Population Statistics for the Village of Kōmi, Motosu District, Mino Province, in the Ogaki Domain) Mita Gakkai Zasshi, Vol. 10, II (1960), pp. 200–201Google Scholar . And in Yokouchi village of Shinano Province, “from 1671–1725 married women who survived until age 45 with their husbands had 6.5 births if the women married between 15–20, 5.9 if married at 21–25, and 1.5 if at over 26. From 1726–75 they averaged 4.2 births if married between 15–20, 3.4 if at 21–25, and 3.0 if at over 26. And after 1776, women averaged 3.6–3.8 births whenever they married.” The English summary appended to Akira, Hayami, “Shūmon-aratame-chō o tsūjite mita Shinshū Yokouchi-mura no chōki jinkō tōkei—Kambun 11-Meiji 4” (Long-term Population Statistics of the Village of Yokouchi of Shinano Province Seen Through The Religious Investigation Registers, 1671–1871), Sangyō Kenkyūsho Shirizu No. 202 (Management and Labor Study Series No. 202) (Tokyo: Keiō Gijuku Daigaku Sangyō Kenkyūsho, 1967–68)Google Scholar . These data are obviously not directly comparable, but it can easily be seen that the Fujito averages are on the low side. Kandoshinden was a village located in a reclaimed area and thus may be considered to have been relatively underpopulated in the Tokugawa period, by comparison to earlier settled areas.

²⁷ Minoru, Nishijima, Edo jidai no sei seikatsu (Sex Life in The Edo Period) (Tokyo: Yūzankaku, 1969), pp. 241–242Google Scholar.

²⁸ Sumio, Taniguchi, Okayama hanseishi no kenkyū (A Study of the History of the Domain of Okayama) (Tokyo: Hanawa Shobō, 1964), p. 503 ffGoogle Scholar . The domain tried to regulate the establishment of branch families for fear that the unregulated establishment of new households would weaken the tax base.

²⁹ In a sample analysis made on 11 years between 1775 and 1863, the percentage of persons classified sister-in-law to those classified brothers ranged from a low of 27.3 percent in 1810 to a high of 54.8 percent in 1833. The years in which the percentage was highest were 1833 and 1841, while the years in which it was lowest were 1810, 1856, and 1863.

³⁰ Iinkai, Fujito Chōshi Henshū, ed. (Editorial Board of “A History of the Town of Fujito”), Fujito chōshi (A History of the Town of Fujito) (Okayama: Fujito Chōshi Henshū Iinkai, 1955), p. 163Google Scholar.

³¹ These figures were obtained from the Kojima gunson mura meisai-chō (Detailed Village Records on Villages in The District of Kojima), collected and copied by Nagayama Usaburō and now located in the Kurashiki Municipal Library. Tax assessments, as well as holdings of individual farmers, were all calculated on the basis of output in rice equivalent in the koku unit (1 koku equals 5.12 American bushels) rather than on land area. Since the productivity of land varied widely, output rather than area better approximates wealth.

³² Fujito chōshi, pp. 120–121. The percentage is an estimate made by the author of this article on the basis of a map on p. 120. No data were provided.

³³ These figures, compiled from the O'nengumai toritate sanyō-chō (Records of Rice Tax Computations) of Fujito are shown graphically in Ota Ken'ichi and Matsuo Keiko, “Bakumatsu-Meiji shoki ni okeru jinushisei no tenkai—Okayama-han Kojima no kōri Hikasa-ke o chūshin to shite” (The Development of the Landlord System in the Bakumatsu and Early Meiji Periods: With a Focus on the Hikasa House of the Kojima District in the Domain of Okayama), Okayama Shigaku (Okayama History), No. 7–6 (June 1960), p. 63. The authors of the Okayama-ken no rekishi (The History of Okayama Prefecture) (Okayama: Okayama-ken, 1962) stated: “That such large landholders as the Hikasa could increase their economic power was due to the fact that the land tax [nengu] did not rise part passu with agricultural productivity. This meant that the land tax was low and the tenants worked efficiently on their tenant farms,” p. 374.

³⁴ Ota and Matsuo, p. 52.

³⁵ Ota and Matsuo, p. 76.

³⁶ Information on the development of the production of Kogura-ori was obtained largely from Ono Masao, “Okayama-han ni okeru Ogura orimono no ryūtsū keitai—Kaei-Ansei-ki o chūshin to shite” (Patterns of Trade in Kogura Weave in the Domain of Okayama: With a Focus on the Period 1848–1854), in Hōgetsu Keigo Sensei Kanreki Kinenkai, ed. (Committee for the Commemoration of the Sixty-first Birthday of ProfessorKeigo, Hōgetsu), Nihon shakai keizaishi kenkyū, Kinsei hen (A Study of the Social and Economic History of Japan, Volume on the Tokugawa Period) (Tokyo: Yoshikawa Kōbunkan, 1967), pp. 440–465Google Scholar , and from Taniguchi Sumio, Okayama hanseishi no kenkyū, pp. 631–634.

³⁷ Ono, p. 460.

³⁸ Shiyakusho, Okayama, Okayama shishi, Sangyō keizai hen (The History of Okayama City, Volume on Industry and the Economy) (Okayama: Okayama Shiyakusho, 1966), p. 286Google Scholar . Around 1850, Hikasa Yūtarō was also authorized to sell Kogura-ori to engoku, “distant places.” Ono, p. 454.

³⁹ Sumio, Taniguchi, “Kinsei ni okeru Bizen Minami Kojima no shōhin seisan to ryūtsū” (Commodity Production and Distribution in Southern Kojima of Bizen in the Tokugawa Period), in Takeichirō, Fukuo, ed., Naikai sangyō to suiun no shiteki kenkyū (Historical studies of Industry and Water Transport in the Inland Sea) (Tokyo: Yoshikawa Kōbunkan, 1966), p. 265Google Scholar . That such by-employments were common in Tokugawa Japan can be seen in Smith, Thomas C., “Farm Family By-employments in Preindustrial Japan,” The Journal of Economic History, Vol. XXIX, No. 4 (Dec. 1969), pp. 687–715CrossRefGoogle Scholar.

⁴⁰ This incident is cited in Ono, p. 460.

⁴¹ This undoubtedly reflects the declining role of the castle town as a commercial center.

⁴² The cotton industry in particular attracted labor as wages in this industry were high. Taniguchi, Okayama hanseishi no kenkyū, p. 632.

⁴³ Ota and Matsuo, p. 85. These people came as tenant farmers, but many were later able to acquire land of their own.

⁴⁴ Taniguchi discusses this in “Kinsei ni okeru Bizen Minami Kojima no shōhin seisan to ryūtsū,” p. 264 ff.

⁴⁵ See Spengler, Joseph J., “Demographic Factors and Early Modern Economic Development,” Daedalus (Spring 1968), pp. 441–444Google Scholar . The percentage of the population aged 15–64 in northern European countries in the eighteenth and nineteenth centuries tended to be around 60, a percentage which “was favorable to average productivity,” p. 441. In contrast, the countries of Africa, Asia, and Latin America today have percentages well below 60.

⁴⁶ Seiichi, Andō, Kinsei zaikata shōgyō no kenkyū (A Study of Commerce in the Rural Areas in the Tokugawa Period) (Tokyo: Yoshikawa Kōbunkan, 1958), p. 125Google Scholar . The domain government felt the peasants were overspending on items they did not need, evidence in itself that the peddlars must have been selling items other than sheer necessities.

⁴⁷ Fujisawa Yasushi, “Shōhin ryūtsū kara mita jōkamachi Okayama to zai” (The Castle Town of Okayama and the Countryside Seen from Commodity Distribution), Okayama-ken no Rekishi (History of Okayama Prefecture) (journal, n.d.), p. 173.

⁴⁸ Andō, p. 129.

⁴⁹ Age at marriage and first birth, average number of children in the completed family, the interval between the end of childbearing and the termination of the marriage or age 45, the sex ratio of children born, and the incidence of adoption were all analyzed by the landholding class of women married between 1825 and 1841 and women married after 1841 but whose marriages and/or childbearing years were terminated prior to 1863.

⁵⁰ The sample comprised 100 women. Even a household holding 3 to 6 tan (¾ to 1½ acres) in land could by no means be considered a large landowner. Undoubtedly these households too engaged in by-employments in order to make what was considered an adequate income.

⁵¹ For a list of available population figures on Kojima and the other districts in Bizen Province of Okayama, see Hanley, Susan B., “Population Trends and Economic Development in Tokugawa Japan: The Case of Bizen Province in Okayama,” Daedalus (Spring 1968), p. 627Google Scholar.

⁵² A disclaimer should be made here as to what is meant by rise in standard of living. It is not argued that Japanese peasants were living well by today's standards—in fact, it is not hard to find evidence on the poor living conditions in many villages—but that in relative terms the people in Fujito were better off in the mid-nineteenth century than they were a century or two earlier.

⁵³ Examples of such customs are a high average age at marriage for women and low proportions married. See J. Hajnal, European Marriage Patterns in Perspective,” Population in History, pp. 101–143. These customs practiced in western Europe are considered by Hajnal to be “unique or almost unique in the world.” p. 101.

⁵⁴ This hypothesis is expounded in Deane, Phyllis and Cole, W. A., British Economic Growth, 1688–1959, Second Edition (Cambridge: Cambridge University Press, 1967), pp. 96–97Google Scholar.

⁵⁵ Wrigley, E. A., “Family Limitation in Pre-Industrial England,” Economic History Review, 2nd Series, Vol. 19, No. 1 (1966), p. 109CrossRefGoogle ScholarPubMed.